Testing for Unit Roots: Bayesian Approach Versus Classical Tests: An Application to the Turkish Economy

1993
Selçuk, Faruk
This paper investigates time series properties of major macroeconomic series in the Turkish economy. Particularly, we are interested in whether the series are stationary in levels, possibly with a trend, or stationary in the first differences. For testing unit roots, it is known that the classical procedure has tow power against some specific alternatives. Recently, Sims (1988) proposed a Bayesian test to discriminate between a unit root and a root close to but below unity. In this paper, both tests are applied to some major macroeconomic variables. While the results of a classical test indicate the existence of a unit root for all of the series except one, the Sims test rejects the unit root hypothesis for some of these series. According to the Sims test, the existence of random walk in these series requires a strong prior probability on the unit root hypothesis. The results of the study suggest that relying upon only classical test results may lead to specification errors in applied work.
Citation Formats
F. Selçuk, “Testing for Unit Roots: Bayesian Approach Versus Classical Tests: An Application to the Turkish Economy,” ODTÜ Gelişme Dergisi, vol. 20, no. 3, pp. 357–379, 1993, Accessed: 00, 2024. [Online]. Available: https://hdl.handle.net/11511/112299.